create-objectstack 17.0.0-rc.6 → 17.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/CHANGELOG.md +190 -0
  2. package/package.json +1 -1
package/CHANGELOG.md CHANGED
@@ -1,5 +1,195 @@
1
1
  # create-objectstack
2
2
 
3
+ ## 17.0.0
4
+
5
+ ### Major Changes
6
+
7
+ - e47b342: feat!: require Node.js 22 — promise the runtime we actually test (#3825)
8
+
9
+ Every published package declared `engines.node: ">=18.0.0"`. **Node 18 reached
10
+ end-of-life on 2025-04-30 and Node 20 on 2026-04-30**, so the compatibility
11
+ promise covered two runtimes nobody patches — and, after #3830 moved CI to Node
12
+ 22, two runtimes nothing in this repo verifies.
13
+
14
+ That left the promise and the evidence with **no overlap at all**:
15
+
16
+ | | Node version |
17
+ | ----------------------------------------------------------------------------------------------- | ------------ |
18
+ | What CI validates every PR on | **22** |
19
+ | What `release.yml` publishes from | **22** |
20
+ | What every shipped Docker image runs (`docker/Dockerfile`, `blank` template, self-hosting docs) | **22** |
21
+ | What `engines.node` promised users | **>=18** |
22
+
23
+ `engines.node` is now `>=22.0.0` across all 50 manifests. This is the honest
24
+ floor: it is the only runtime the packages are built, tested and shipped on.
25
+
26
+ ## Migration
27
+
28
+ **If you are on Node 22 or newer, nothing changes.** Node 24 (Active LTS since
29
+ 2025-10-28) and Node 26 both satisfy the new range.
30
+
31
+ If you are on Node 18 or 20, upgrade to Node 22+. Both are past end-of-life and
32
+ receive no security patches:
33
+
34
+ ```bash
35
+ nvm install 22 && nvm use 22
36
+ ```
37
+
38
+ npm and pnpm surface an unsatisfied `engines` as an **`EBADENGINE` warning**, not
39
+ a hard failure, so an existing install will not break the moment you upgrade —
40
+ but the package is no longer tested on that runtime, and the failures are the
41
+ kind that do not announce themselves. #3812 is the worked example: a native
42
+ dependency whose `engines` required a newer Node loaded anyway on the older one
43
+ and then killed the test worker at the process level, with no JS error and a
44
+ summary that still said "passed".
45
+
46
+ If your CI pins Node, pin it to 22 as well — running your gates on a runtime
47
+ your dependencies no longer support is exactly the split this change closes.
48
+
49
+ ## Also updated
50
+
51
+ The "Node 18+" prerequisite was restated in ten user-facing places
52
+ (`README.md`, `CONTRIBUTING.md`, the getting-started and deployment docs, the
53
+ todo example, and the `objectstack-platform` skill's `compatibility` field).
54
+ All now say 22. Changelogs and ADRs are historical records and were left alone.
55
+
56
+ ### Patch Changes
57
+
58
+ - 9f060e5: chore(deps)!: better-auth 1.7.0-rc.2 (account identity restructuring) + the
59
+ production-dependency batch from #3517
60
+
61
+ **better-auth 1.7.0-rc.1 → 1.7.0-rc.2** across the family (`better-auth`,
62
+ `@better-auth/core`, `@better-auth/oauth-provider`, `@better-auth/sso`, and the
63
+ adapter/telemetry overrides). `@better-auth/scim` deliberately stays on
64
+ 1.7.0-rc.1 — rc.2 replaces its whole model (code-defined connections; the
65
+ `scimProvider` model and the generate-token endpoint are gone), which is a
66
+ feature migration, not a version bump. Its peer range accepts rc.2 core, and the
67
+ advisory that forced the original pin (GHSA-j8v8-g9cx-5qf4) is still fixed.
68
+
69
+ **BREAKING — account identity.** better-auth renamed `account.accountId` to
70
+ `account.providerAccountId` and added a REQUIRED `account.issuer`; sign-in now
71
+ resolves accounts by `(issuer, providerAccountId)`.
72
+
73
+ - FROM `fields: { accountId: 'account_id' }` → TO
74
+ `fields: { issuer: 'issuer', providerAccountId: 'account_id' }`. The provider
75
+ account id keeps its `account_id` column — only the better-auth-side name
76
+ moved — and `sys_account` gains an `issuer` column.
77
+ - FROM `internalAdapter.createAccount({ providerId, accountId, … })` → TO
78
+ `createAccount({ providerId, issuer, providerAccountId, … })`. A local
79
+ password account carries the issuer better-auth mints for itself,
80
+ `local:credential`.
81
+ - FROM `client.auth.accounts.unlink({ providerId, accountId })` → TO
82
+ `unlink({ accountId })`, where `accountId` is now the account ROW id (the `id`
83
+ from `accounts.list()`), matching better-auth's narrowed body.
84
+ `accounts.list()` returns `issuer` + `providerAccountId` in place of
85
+ `accountId`.
86
+
87
+ **Existing deployments:** rows written before 1.7 have no issuer and are
88
+ invisible to sign-in until stamped. The auth plugin now runs an idempotent
89
+ boot-time backfill that stamps what it can derive — `local:credential` for
90
+ password accounts, `local:oauth:<providerId>` for configured social providers,
91
+ and the registered IdP's real `iss` from `sys_sso_provider` for federated ones.
92
+ Accounts from a federated IdP that is no longer registered cannot be derived;
93
+ they are logged with their provider id and row count rather than guessed, and
94
+ those users cannot sign in through that provider until the row is stamped with
95
+ the IdP's issuer or removed so a fresh login re-links it.
96
+
97
+ **Also required by 1.7:** `SecondaryStorage` gained two mandatory methods, both
98
+ now implemented over the kernel cache service — `getAndDelete` (single-use
99
+ verification values) and `increment` (fixed-window rate-limit counter;
100
+ `rateLimit.storage: 'secondary-storage'` throws at boot without it).
101
+
102
+ The rest of #3517's production-dependency batch rides along: `@oclif/core`
103
+ 4.13.0, `@hono/node-server` 2.0.12, `hono` 4.12.32, `tar` 7.5.22, `jose` 6.2.4,
104
+ `pinyin-pro` 3.28.2, plus the private docs app's fumadocs/next/react bumps.
105
+
106
+ - 2e836de: chore(packaging): CHANGELOG.md ships in every npm tarball (#4261)
107
+
108
+ The AGENTS.md post-task checklist requires breaking changesets to carry their
109
+ FROM → TO migration because "this text ships to consumers as `CHANGELOG.md`
110
+ inside the npm package and is what an upgrading agent greps after the tombstone
111
+ error." That delivery path was severed for 68 of the 69 publishable packages:
112
+ npm packs `package.json` / `README*` / `LICENSE*` unconditionally but — unlike
113
+ older npm versions — not `CHANGELOG.md`, and the canonical
114
+ `"files": ["dist", "README.md"]` whitelist never named it. Measured on npm
115
+ 10.9.7: `npm pack --dry-run` on `@objectstack/types` shipped 3 files while its
116
+ 70KB `CHANGELOG.md` stayed behind. Only `@objectstack/spec` listed it
117
+ explicitly.
118
+
119
+ The tombstone-error scenario is precisely the one where the repo is out of
120
+ reach — the upgrading agent has `node_modules` and nothing else — so the
121
+ migration text has to ride in the tarball. Every publishable package now
122
+ declares `CHANGELOG.md` in `files`, and the canonical whitelist is
123
+ `["dist", "README.md", "CHANGELOG.md"]`.
124
+
125
+ The other half is the gate: `check:published-files` gains a fifth invariant,
126
+ COMPLETE — a whitelist that fails to cover `CHANGELOG.md` fails the
127
+ always-required lint job, so the next package cannot silently sever the path
128
+ again. `@objectstack/spec`'s per-package EXTRA_ENTRIES exemption dissolves
129
+ into the canonical set.
130
+
131
+ Consumer-visible change: one more file per install (the package's changelog,
132
+ e.g. 70.8KB for `@objectstack/types`), and `grep -r "removed key"
133
+ node_modules/@objectstack/*/CHANGELOG.md` now finds the migration it was
134
+ promised.
135
+
136
+ - 4e9e184: chore(deps): OSV security batch — bump tar to ^7.5.21 (GHSA-r292-9mhp-454m) and
137
+ js-yaml to ^5.2.2 (GHSA-pm4m-ph32-ghv5)
138
+
139
+ Both are declared-range bumps to the patched releases, so downstream installs
140
+ resolve the fixed versions from the published manifests, not just this
141
+ workspace's lockfile. The same batch clears the remaining transitive advisories
142
+ (next 16.2.11 in apps/docs; workspace overrides for brace-expansion, sharp,
143
+ react-router, @sveltejs/kit, @hono/node-server) — those live in pnpm-workspace.yaml
144
+ and the private docs app, which do not ship.
145
+
146
+ - 8d41998: fix(create-objectstack): scaffolding a remote template no longer produces a project that cannot build (#4926)
147
+
148
+ `npx create-objectstack@latest my-app -t todo` (and `compliance`, `content`,
149
+ `contracts`, `procurement`) generated a project that failed `objectstack build`
150
+ immediately — 5 of the 6 offered templates. Only the bundled `blank` worked.
151
+
152
+ The scaffolder read the template's original namespace from
153
+ `objectstack.manifest.json`, and that filename names two different documents.
154
+ The bundled template's is app-shaped and carries `namespace`; a remote
155
+ template's is the template-registry document
156
+ (`$schema: …/template-manifest.json`) and carries none — its namespace lives
157
+ only in `objectstack.config.ts`. So the value came back `undefined` for every
158
+ remote template and the object-name rewrite was skipped, while the config's
159
+ `namespace:` was rewritten anyway. The result was `namespace: 'my_app'` sitting
160
+ next to `name: 'todo_task'`, which the `${namespace}_${shortName}` rule rejects.
161
+ Across the five templates, 74 object names were left unrewritten.
162
+
163
+ `objectstack.config.ts` is now the authority for the template namespace (it
164
+ holds the very literal the scaffolder overwrites, so the two cannot disagree),
165
+ with the manifest as fallback. The rewrite also verifies itself: any surviving
166
+ stale prefix throws at the scaffold, naming the files and lines, instead of
167
+ surfacing as a build failure on the user's first command.
168
+
169
+ - 7309c81: chore(cli,create-objectstack): scaffolds no longer name a driver (#4065)
170
+
171
+ `os init` and the `create-objectstack` blank template both listed
172
+ `@objectstack/driver-memory` in the generated `dependencies`. It was the only
173
+ driver named, which read as an endorsement — "this is the driver your app runs
174
+ on" — when it is in fact the **last-resort rung** of the dev step-down (native
175
+ `better-sqlite3` → WASM SQLite → mingo). A new project's first impression of the
176
+ data layer should not be the engine that enforces no primary keys, no
177
+ uniqueness, no `NOT NULL` and no column types.
178
+
179
+ It was also redundant: `@objectstack/runtime` already depends on `driver-sql`,
180
+ `driver-sqlite-wasm` and `driver-memory`, and every script in both scaffolds runs
181
+ through the CLI, which carries all four. Removing the line changes nothing a
182
+ generated project can do — `objectstack dev` still resolves SQLite by default,
183
+ and `OS_DATABASE_URL` still selects Postgres / MySQL / MongoDB.
184
+
185
+ Docs updated to match: the "packages you depend on" table in _Your first project_
186
+ no longer lists a driver row (it now says where drivers come from), and the
187
+ Memory Driver section of _Database Drivers_ documents the opt-in persistence
188
+ default, carries a migration callout for the old `'auto'` behaviour, and points
189
+ test authors at in-memory SQLite. That section also claimed "Data is lost when
190
+ the process exits", which was simply false while `'auto'` was the default — it
191
+ wrote a file into the working directory.
192
+
3
193
  ## 17.0.0-rc.6
4
194
 
5
195
  ## 17.0.0-rc.5
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "create-objectstack",
3
- "version": "17.0.0-rc.6",
3
+ "version": "17.0.0",
4
4
  "description": "Create a new ObjectStack project — npx create-objectstack",
5
5
  "bin": {
6
6
  "create-objectstack": "./bin/create-objectstack.js"